How to fit a log-normal dist. to data in x-data and y-data format using 'lognfit'?

Hi,

This is the problem: I have a set of points to which I'd like to fit the log-normal distribution. I have two vectors for their X and Y values. I ran into 'lognfit' function which just accept a single vector as its input. I know this can be performed by something like 'lsqcurvefit', but I need to get the confidence intervals for the mu and sigma which 'lognfit' function returns easily as its output. Could anyone help please?

Answer 1

[Torsten]

http://de.mathworks.com/help/stats/nlparci.html

Best wishes

Torsten.

[Mohammad Monfared]

Thanks a lot!

[Torsten]

If you have the original data from which the (X,Y) vectors have been derived, you should work with them and use "lognfit" to get mu and sigma. I guess the confidence intervals for the parameters will be different if you explicitly know the underlying distribution.

Best wishes

Torsten.